1

(1) I want to do a multivariate regression in R (where each output sample is a vector, instead of a number), which I know can be handled using the lm() function; however, the multivariate output data is not necessarily normally distributed, so I'm trying to use multivariate t regression as an alternative. Does that sound ok?

(2) Is there a package in R to do multivariate student-t regression?

Glen_b
  • 282,281
  • 3
    It's not clear what you mean by "the multivariate output data is not necessarily Normally distributed," or why you think that is important. – EdM Jun 29 '15 at 21:40
  • where each output sample is a vector, instead of a number is not clear? Do you mean continuous variables? R package mvtnorm may be used for multivariate regression. – Dr Nisha Arora Mar 18 '16 at 14:36
  • (2) is off-topic and (1) would probably need more details to be answerable, so voting to close – Juho Kokkala Oct 27 '17 at 19:18

0 Answers0